On Fri, Oct 12, 2018 at 10:04:36AM +0300, Alexey Budankov wrote: > On 11.10.2018 12:29, Peter Zijlstra wrote: > > On Thu, Oct 11, 2018 at 07:50:23AM +0000, Song Liu wrote: > >> Or step back a little... I see two big changes: > >> > >> 1. struct perf_ctx_context is now per cpu (instead of per pmu per cpu); > >> 2. one perf_event_ctxp per task_struct (instead of 2). > > > > Correct, we reduce to 1 cpu context and 1 task context at all times. > > This in fact simplifies quite a bit of things. > > And what is currently missing is > some markup of the per cpu event list into per pmu sublists and > capability to rotate or not rotate the sublists independently, > right?
Yes, that is what the new perf_event_pmu_context is. That tracks the per pmu sublist state.
